Correlation between DNA microsatellite heterozygosity and growth rate was investigated in two hatchery-propagated stocks of the shrimp Penaeus stylirostris, which had been genetically isolated from wild founders for 17 generations, and from each other for five more generations. presumed demographic history of these populations suggested that they were maintained at small effective population sizes. and this was confirmed by significant changes in allelic frequencies between the two stocks. Despite a small sample size (n = 60) and a small number of loci screened, a significant positive correlation was detected between microsatellite tri-locus heterozygosity and growth rate in one of the two stocks (r = 0.3, P = 0.02). In the other stock (n = 48), the positive correlation was not significant, but the combined test of single locus heterozygote advantage over the three loci was significant (P = 0.005).Use of DNA microsatellite markers, whose selectively neutral status is generally accepted, allowed us to reject the hypothesis of direct overdominance at marker loci as an explanation for an association of growth rate with heterozygosity. The average depression associated with one locus in our case is above or equivalent to the amount observed for a 10% increase of inbreeding in other species. These results suggest that heterozygosity at neutral marker loci is sufficiently well correlated with individual inbreeding coefficients to reveal a significant residual inbreeding load for growth rate in Tahitian P. stylirostris stocks. Inbreeding during hatchery propagation was, thus, insufficient to purge the ancestral load.